On invariant measure for semilinear equations with dissipative nonlinearities (Q4029012)

[For the entire collection see Zbl 0754.00020.] The authors establish the theorem on the existence and uniqueness of a ``generalized'' mild solution of semilinear partial differential stochastic equations of the type \[ dX(t)=[AX(t)+F(X(t)]dt+dW(t),\qquad X(0)=\xi, \tag{1} \] where \(A:D(A)\subset H\to H\) is a generator of a semigroup \(S(t)\) and \(F:D(F)\subset H\to H\) is a nonlinear operator which is dissipative in the separable Hilbert space \(H\), \(W(\cdot)\) is an \(H\)- valued cylindrical Wiener process, and \(B\) is a Banach space embedded continuously and densely in \(H\) such that the stochastic convolution \(W_ A(\cdot)\) is a \(B\)-continuous process; here \[ W_ A(t)=\int^ t_ 0 S(t-s)dW(s). \] A ``generalized'' solution of (1) is defined as a process \(X\) with \(X(0)=x\in H\) for which there exist \(\{x_ n\}\subset B\), \(\| x-x_ n\|_ H\to 0\) \((n\to 0)\) such that ``usual'' solutions \(X_ n\) of (1) with \(X_ n(0)=x_ n\in B\) converge to \(X\) in \(C([0,T];H)\) \(P\)-a.s. Under other natural additional conditions the authors prove the existence of a unique invariant measure for the Markov semigroup in \(H\) generated by generalized solutions of (1).